AI services pricing is about to break the day-rate model

The day-rate model has survived every wave of technology services transformation. AI may finally break it.

Globant reported that annual recurring revenue from Glob.AI reached $52.8 million during the second quarter of 2026, up 61% quarter on quarter, and expects at least $110 million by year end. The company offers supervised AI Pods that complete defined work using AI agents and human specialists. Customers pay for an output or measured consumption rather than the hours and seats behind it.

These are Globant’s own figures, not independent proof of a market-wide shift. They are still a meaningful commercial signal: enterprises appear willing to buy technology services through a model that does not begin with a team structure and rate card.

Traditional services pricing makes effort visible. Procurement negotiates rates, seniority mix, expenses, and change control. The customer often carries productivity risk.

Output-based AI services allow a supplier to use any mix of agents, models, automation and human supervision to produce an agreed result. It replaces the question “how many people?” with “what exactly counts as complete?”

That requires specific definitions: quality threshold, acceptance criteria, independent measurement of consumption, included retries and rework, treatment of model changes, accountability for failed outputs, data and IP boundaries, and a clear distinction between changing requirements and supplier failure.

Compare against the equivalent human team, but do not let that become the sole pricing mechanism. The stronger negotiation focuses on measurable volume, quality, risk allocation, and value. Contracts should also share productivity gains as the delivery model improves over time.

The next services negotiation will be won by the team that defines the output best.

Source: Globant Q2 2026 results